Teen relationship violence (TRV) can pose immediate and serious threats to a teen's physical and mental health. Many high schools offer violence prevention programs, but few have been rigorously evaluated, and of the few that have demonstrated positive effects in reducing rates of violence, most are lengthy and/or require highly trained facilitators. In short, there is a need to develop cost-effective methods for preventing TRV. The proposed research will evaluate an innovative online bystander intervention, which has the potential to be much more cost-effective than current programs designed to prevent TRV. We will recruit 240 high-school students, grades 9-12, who will be randomly assigned to the online bystander intervention condition or to an online control condition. There will be 3 primary assessment points (baseline, post-treatment, 6-month follow-up). In addition to using standard methods of assessing bystander behavior, the proposed research will use an innovative observational assessment of bystander behavior that makes use of virtual reality. We believe this new method has the potential to offer stronger evidence of program efficacy than self-report methods alone. Furthermore, the proposed research is designed to gain insights into mediators and moderators of treatment effects. If the online intervention increases bystander behavior in a manner that deters TRV, or if it reduces teens' TRV perpetration, this research will lay the groundwork for a large-scale evaluation of the intervention. An effective online program would have enormous potential in efforts to reduce TRV, because it could be implemented broadly and cost-effectively.